IEPC M. Bodendorfer 1, K. Altwegg 2 and P. Wurz 3 University of Bern, 3012 Bern, Switzerland. and

Size: px

Start display at page:

Download "IEPC M. Bodendorfer 1, K. Altwegg 2 and P. Wurz 3 University of Bern, 3012 Bern, Switzerland. and"

Clare Leonard
5 years ago
Views:

1 Future thruster application: combination of numerical simulation of ECR zone and plasma X-ray Bremsstrahlung measurement of the SWISSCASE ECR ion source IEPC Presented at the 31st International Electric Propulsion Conference, University of Michigan Ann Arbor, Michigan USA M. Bodendorfer 1, K. Altwegg 2 and P. Wurz 3 University of Bern, 3012 Bern, Switzerland and H. Shea 4 Ecole Polytechnique Fédérale de Lausanne, 2002 Neuchatel, Switzerland Abstract We present the combination of the numerical 3D magnetic field simulation of SWISSCASE, the new 10 GHz ECR ion source at the University of Bern in Switzerland, and the experimental X-ray Bremsstrahlung measurement of the same ion source, to determine the hot electron and the total electron number densities of its ECR plasma. The results, which are in excellent agreement with literature, demonstrate the quality of the numerical simulation, the X-ray Bremsstrahlung measurement and the method of combination to calculate these particle densities. The revealed key parameters of the ECR plasma such as geometry, particle densities and Bremsstrahlung emission, may enable the development of a microwave ECR plasma model, yet to be defined, which can be integrated into a Finite Difference Time Domain (FDTD) simulation to reproduce realistic microwave-plasma interaction. The presented method of 3D magnetic field modeling, X-ray Bremsstrahlung measurement and the subsequent determination of the electron densities is not limited to the SWISSCASE ion source but can further be used for detailed investigation and optimization of current and future electric propulsion concepts. ω MW = microwave angular frequency, Hz B = magnetic field density, T m e = electron mass, kg e = electron charge, C n ehot = hot electron number density, 1/m 3 n e = total electron number density, 1/m 3 n z = particle density of ion charge state z, 1/m 3 Φ = plasma potential, V V = volume, m 3 T = temperature, ev Nomenclature 1 Postdoctoral researcher, Space Research and Planetary Sciences, michaelbodendorfer@space.unibe.ch; present address: ISAS/JAXA, Japan, bodendorfer@ep.isas.jaxa.jp 2 Professor, Space Research and Planetary Sciences, kathrin.altwegg@phim.unibe.ch 3 Professor, Space Research and Planetary Sciences, peter.wurz@space.unibe.ch 4 Professor, Microsystems for Space Technologies Laboratory, herbert.shea@epfl.ch 1

2 Bodendorfer et al. 1 THE SIMULATED PARTICLE CONFINEMENT AND BREMSSTRAHLUNG (a) (b) Figure 1: (a) Cross sections of the SWISSCASE permanent magnet confinement (Pink: ECR plasma) along the main axis and (b) detail of the Halbach hexapole confinement perpendicular to the main axis. Arrows in the permanent magnetic material indicate the direction of magnetization. All measures in mm. f f /c = hot electron fraction A = area, m 2 r m = magnetic confinement mirror ratio r pl = ECR zone radius, m t = time, s E = energy, ev or J Introduction Inside an ECR ion thruster, the ECR plasma is an active microwave element. Therefore, detailed knowledge about the locality of the ECR zone is required to precisely model the microwave situation inside such a thruster. In addition, knowledge of the magnetic field is required for the ion optical simulation of Hall, DC, Ion and MPD thrusters to ensure thruster efficiency and to minimize sputtering of electrodes and acceleration grids. We demonstrate the precision of the 3D simulation of SWISSCASE, the Solar Wind Ion Source Simulator for the CAlibration of Space Experiments, a new 10 GHz min-b ECR ion source, installed at the University of Bern in Switzerland. The obtained precision in the simulation of complex 3D magnetic fields can be used not only for the design of ion sources but also for the design of electric propulsion systems. Generally the determination of the hot electron density inside an ECR plasma by probes is hindered and limited by multiple plasma-probe-interaction effects. We present the determination of the electron density inside SWISSCASE using the numerical simulation of its confinement in combination with an X-ray Bremsstrahlung measurement of SWISSCASE. Comparing the results with the cut-off limited electron density demonstrates the usability of the numerical simulation and the method of the electron density calculation. 1 The simulated particle confinement and Bremsstrahlung The magnetic plasma confinement of SWISSCASE is formed by permanent magnets and will be used to produce ions for the calibration of mass spectrometers such as ROSINA 6 flown aboard the space-craft ROSETTA and PLASTIC 7 flown aboard the space-craft STEREO. 6 Rosetta Orbiter Spectrometer for Ion and Neutral Analysis 7 PLAsma and Supra-Thermal Ion Composition investigation 2/9

3 Bodendorfer et al. 2 THE X-RAY MEASUREMENT Two cross sections of the simulated particle confinement of SWISSCASE are shown in Fig.1. Using Opera3D by Vectorfields, a comprehensive FEM simulation of the magnetic confinement field, the plasma potential as well as ion and electron trajectories of the complex Halbach hexapole [14] responsible for the radial particle confinement was performed [4]. Fig.2 shows the simulated magnetic field along the main axis of SWISSCASE and compares it to the measured value. Figure 2: SWISSCASE: Comparison of measured and simulated magnetic field on the main axis of SWISSCASE. The maximum relative error at the ECR zone is below 1%. In the 3D model, the ECR zone represents a volume enclosed by an iso surface of constant magnetic field, which fulfills the ECR criteria (see Eq.2). Eq.2 allows to determine the magnetic field which qualifies for the ECR mechanism [1]. B ECR = ω MW m e e All units are in mks, with B ECR as the magnetic field required for resonance, ω MW the incident microwave angular frequency, m e the electron mass and e as its charge. For SWISSCASE the required magnetic field is mt. We used this value to select the elements in the FEM simulation result to represent the ECR zone and to precisely determine its geometry and locality. Fig.3a and 3b show a section view of the 3D magnetic simulation result. We can identify the iso-contour line which is qualified for the ECR mechanism by its magnetic field value of 388 mt. To further help visualizing the 3D structure of the ECR zone, Fig.4 shows a 3D view of the simulated SWISSCASE ECR zone with section cut. The simulation resulted in precise geometrical data about the ECR zone of SWISSCASE. Table 1 gives a summary of its geometrical properties which we used for all further calculations of the Bremsstrahlung radiation power density and the electron number densities. ECR zone diameter 12 mm ECR zone length 13 mm ECR zone volume 980 mm 3 ECR zone surface area 478 mm 2 (1) Table 1: Summary of geometrical properties of the SWISSCASE ECR zone. 2 The X-ray measurement The SWISSCASE ECR plasma features at least two electron fractions, one with a low temperature, in thermodynamic equilibrium with the bulk plasma and another one, a hot fraction, with a higher temperature [4], which dominates the ionization process. The hot electron fraction produces a Bremsstrahlung spectrum by collisions with ions and neutrals. We used this Bremsstrahlung spectrum to determine the temperature of the hot electron population. Fig.6 shows a cross section of the X-ray measurement setup which is used to produce the following Bremsstrahlung spectra. 3/9

result and (b) detailed view of the center part of (a).

Figure 4: The ECR zone of SWISSCASE: Isosurface of a constant magnetic field (left image) and cut-away seen

Each ball represents one element of the FEM simulation qualified by the ECR criteria B = ω m e /e = 388.7 mt.

4 Bodendorfer et al. 2 THE X-RAY MEASUREMENT (a) (b) Figure 3: (a) Magnetic field in section view of the SWISSCASE FEM simulation result and (b) detailed view of the center part of (a). The ECR zone is visualized by a white iso-contour line of 388 mt. Figure 4: The ECR zone of SWISSCASE: Isosurface of a constant magnetic field (left image) and cut-away seen from the negative z-direction (image to the right). Each ball represents one element of the FEM simulation qualified by the ECR criteria B = ω m e /e = mt. The balls at the cutting faces are colored red for better visualization. The diameter of the structure measures 12 mm, its length in z-direction is 13 mm. Figure 5: Setup of the presented X-ray Bremsstrahlung measurement. In light blue, a lead collimator is shown which inhibits secondary X-ray emission from electrons and ions hitting the facility walls. 4/9

5 Bodendorfer et al. 3 THE HOT ELECTRON DENSITY IN THE SWISSCASE ECR ZONE Assuming a Maxwellian kinetic energy distribution for the hot electrons, the intensity of the emitted Bremsstrahlung as a function of the photon energy E photon can be written as [8, 9, 10]: ( I(E photon ) = I 0 exp E ) photon T ehot (2) With I(E photon ) the intensity of the incident photons, I 0 the maximum photon intensity at the minimal photon energy in the exponential decay spectrum and T ehot the hot electron temperature. This rationale shows that an X-ray Bremsstrahlung spectrum obtained from the hot electron population should feature an exponential decay with increasing photon energy. Figure 6: Bremsstrahlung spectrum of two different SWISSCASE argon ECR plasmas. One is obtained at a feed gas pressure of mbar, shown in blue. The other is obtained at a feed gas pressure of mbar, shown in green.the excitation microwave frequency was GHz and the microwave input power was 95 W for both spectra. Two exponential decay functions (red lines) represent electron temperatures of 9 kev for the data in blue and 11 kev for the data in green respectively. Fig.6 shows two spectra obtained from two different SWISSCASE ECR plasmas[5]. The exponential decay fit of 9 kev corresponds to a plasma X-ray measurement with a neutral gas pressure of mbar. Another plasma X-ray measurement with a neutral gas pressure of mbar can be fit with a 11 kev exponential decay for photon energies larger than 35 kev. This finding is in excellent agreement with the fact that the 9 kev plasma features more overall Bremsstrahlung emissions and a slightly lower mean charge state at more overall ion current in the extracted ion beam compared to the 11 kev plasma of SWISSCASE. In the following we will use the 9 kev hot electron temperature to calculate the hot electron number density, because this plasma represents the very well characterized standard working condition of SWISSCASE. In contrast, the 11 kev ECR argon plasma is inferior in ion beam performance for all useful charge state up to Ar The hot electron density in the SWISSCASE ECR zone We use a semi empirical approach by Huba et al. 2007, [3] to determine the Bremsstrahlung power density of the SWISSCASE hot electron population (Eq.3): P Br /V = n ehot T z max e n 1 z 2 n z (3) i=1 5/9

6 Bodendorfer et al. 3 THE HOT ELECTRON DENSITY IN THE SWISSCASE ECR ZONE All units are in mks except T e in ev, with P Br /V the Bremsstrahlung power density, n ehot the electron density, T e the hot electron temperature, n 1 the particle density of charge state one, z the charge state of ion species and n z the ion density of charge state z. The numerical simulation of the magnetic field is used to determine the ECR zone volume V. P Br /V can also be expressed by the integration of the measured X-ray photons incident on the X-ray detector of our measurement (Eq.4): 1 1 P Br /V = K G t obs V plasma Z 0 I(E photon ) de photon (4) with K G as a geometrical factor, extrapolating the Bremsstrahlung radiation incident on the detector to the total Bremsstrahlung radiated by the ECR plasma. Assuming an isotropic Bremsstrahlung radiation pattern of the ECR plasma, K G can be represented by Eq.5: K G = 4 π L2 plasma det A det (5) All units are in mks, P Br /V I(E photon ) is the Bremsstrahlung emission power density of the plasma under investigation, L plasma det is the distance from the plasma center to the detector, A det is the active detector surface, t obs is the time of observation, V plasma is the plasma volume and I(E photon ) is the measured photon intensity count rate which is corrected by the detector efficiency and the X-ray transmission loss of the vacuum window used in the experimental setup. Furthermore, we can assume quasi charge neutrality of the ECR plasma (Eq.6): z max n ehot = n 1 f h/e z nz (6) n 1 All units are in mks, with f h/e = n ehot /n e as the hot electron fraction, z max the maximum charge state of interest in the plasma under investigation and z the charge of the respective ion counted in electron charges. From Eq.3, Eq.4 and Eq.6 we derive Eq.7: z=1 with K I as a plasma parameter ratio represented by Eq.8: n 2 ehot = K G I0 e T e t obs V f h/e K I (7) K I = z max z=0 z nz n 1 z max z 2 nz z=0 n 1 (8) In order to determine K I the ion charge state distribution inside the plasma under investigation is necessary. The ion charge state distribution inside the SWISSCASE ECR plasma is accessible by the charge state distribution of the extracted ion beam, corrected by an extraction function. The extraction function considers that higher charged ions feature a higher charge-to-mass ratio and are better confined by both, the magnetic and the electrostatic confinement[1, 2, 5, 7, 11], provided by the hot, well confined, ECR electrons. The extraction correction function is given by Eq.9[2]: ( mi n z I z rm e T e r 2 exp z Φ ) pl T i (9) with n z the particle density of ion charge state z, m i the ion mass, T e the cold electron temperature, r m the mirror ratio of the magnetic confinement, r pl the plasma radius, Φ the plasma potential and T i the ion temperature. From Eq.9 it is clear that higher charged ions are confined exponentially better than lower charged ions. This leads to a lower mean charge state in the extracted ion beam than in the ECR plasma under investigation. Table 2 gives 6/9

7 Bodendorfer et al. 4 DISCUSSION Charge state Ion beam Ion beam ECR plasma absolute I z in µa relative I z /I 1 relative n z /n 1 Ar Ar Ar Ar Ar Ar Ar Ar Ar Ar Ar Ar Table 2: Summary of the extracted and the actual charge state distribution in a SWISSCASE argon ECR plasma with a hot electron temperature of 9 kev. a summary of the extracted and the actual charge state distribution in a SWISSCASE argon ECR plasma with a hot electron temperature of 9 kev. From Table 2 we can see that the mean charge state in the plasma is higher than in the extracted ion beam, as it is expected from the better confinement of the highly charged ion species. With this charge state distribution we can calculate the plasma parameter ratio K I and consequently the hot electron temperature by Eq.7. Table 3 gives a summary of the calculation parameters and the resulting electron densities for the 9 kev plasma. Parameter Value Neutral pressure mbars X-ray photon count rate I cts/s Plasma Bremsstrahlung total power W Plasma Bremsstrahlung power density W/m 3 Geometry factor K G (Eq.5) Plasma parameter ratio K I (Eq.8, Table 2) Hot electron density /m 3 Hot electron fraction (estimation, see section discussion) 10 % Total electron density /m 3 Cut-off density /m 3 Table 3: Summary and comparison of plasma parameters for a 9 kev SWISSCASE ECR plasma. 4 Discussion The parameter f h/e, the fraction of the hot electron number density relative to the total electron number density inside an ECR plasma, is subject of investigation in ECR research and is very difficult to access experimentally and is expected to vary depending on experimental conditions. Literature values range from f h/e = (Golubev et al., 2000, 37.5 GHz, T ehot = 710 kev)[12] through f h/e = 0.1 (Girard et al., 1994, 10 GHz to 18 GHz, T ehot = 80 kev)[9] up to f h/e = 0.5 (Petty et al., 1991, 10.5 GHz, T ehot = 300 kev)[13]. In the presented calculation method, f h/e is the only parameter which comes with such a large literature range. The hot electron ratio has significant influence on the resulting hot electron and total electron number density as shown in Fig.7, fixing a1ll other calculation input parameters to the values shown in Table 3. Fig.7 shows, that the hot electron and the total electron number number density depend on the hot electron fraction in inverse manner leading to a decreasing total electron number density and an increasing hot electron number density with an increasing hot electron fraction. The total electron number density exceeds the critical cut-off density for a hot electron fraction lower than 1 %. Since we did not observe any significant noise neither in the extracted SWISSCASE ion beam nor in the reflected microwave power, we can exclude an unstable, noisy and over-dense plasma[2]. Hence, the hot electron fraction in the SWISSCASE ECR plasma is higher than 1 %. This 7/9

8 Bodendorfer et al. REFERENCES Figure 7: The resulting hot electron (in red) and total total electron number density (in blue) as a function of the hot electron fraction f h/e from 0.01 to 0.5 with all remaining input parameters fixed to the values shown in Table 3. For comparison, the critical cut-off electron density for an excitation frequency of GHz is shown in green. represents another anchor point for f h/e in current ECR research excluding very low electron fractions below 1 % for SWISSCASE. Regarding the high argon charge states, which SWISSCASE is able to produce, we chose a hot electron fraction of 10 % which is well within the literature value range [2, 6] for a 10 GHz ECR ion source. 5 Conclusion The results of the numeric SWISSCASE ECR zone simulation and the X-ray Bremsstrahlung measurement of the same ion source can be combined to obtain values for hot electron and total electron number densities which are in excellent agreement with literature. Furthermore the presented method to calculate these electron densities reveals further constraints for the value of the hot electron fraction versus the total electron number density, a value hardly accessible by experiment and not yet conclusive in literature. The presented set of plasma parameters is self consistent. If a reliable value for the hot electron fraction f h/e is available for the ECR plasma of interest, the resulting electron density can be used for the development of a microwave-ecr-plasma model for the integration into a Finite Difference Time Domain microwave simulation. Acknowledgements The authors gratefully acknowledge the financial support by the Swiss National Science Foundation. Grant number: / 1 References [1] Chen, F. F.; Introduction to Plasma Physics and Controlled Fusion, Volume 1: Plasma Physics, Second Edition, Plenum Press, [2] Geller, R.; Electron Cyclotron Ion Sources and ECR Plasmas, Institute of Physics Publishing Bristol and Philadelphia, [3] Huba, J. D.; NRL PLASMA FORMULARY 2007, Beam Physics Branch, Plasma Physics Devision, Naval Research Laboratory, Washington, DC [4] Bodendorfer, M., et al.; Identification of the ECR zone in the SWISSCASE ECR ion source, Nuclear Instruments and Methods in Physics Research B, Volume 266, Issue 21, November 2008, pages [5] Bodendorfer, M.; SWISSCASE, Ph.D. thesis, Ecole Polytechnique de Lausanne, Switzerland, September 2008 [6] Hohl, M., et al.; Investigation of the density and temperature of electrons in a compact 2.45 GHz electron cyclotron resonance ion source plasma by x-ray measurements, PLASMA SOURCES SCIENCE AND TECHNOLOGY, Number 14, 2005, pages /9

9 Bodendorfer et al. REFERENCES [7] Hohl, M.; Design, setup, characterization and operation of an improved calibration facility for solar plasma instrumentation, Ph.D. thesis, University of Bern, Switzerland, 2002 [8] Bernhardi, K., et al., X-ray Bremsstrahlung measurements on an ECR discharge in a magnetic mirror, Plasma Physics 24, pages , [9] Girard A., et al. The Quadrumafios electron cyclotron resonance ion source: presentation and analysis of the results Rev. Sci. Instrum. 65, pages , [10] Hutchinson, I. H., Principles of Plasma Diagnostics, Cambridge University Press, [11] Pastukhov, V. P., Classical longitudinal plasma losses from open adiabatic traps, in: Reviews of Plasma Physics, v. 13, Consultants Bureau, New York, pages , [12] Gobulev, S. V., et al. Formation of multi-charged ions and plasma stabilityat quasigasdynamic plasma confinement in a mirror magnetic trap, Rev. Sci. Instrum. 71, pages [13] Petty, C. C., et al. Confinement of multiply charged ions in a cyclotron resonance heated mirror plasma. Phys. Fluids B 3, pages , 1991 [14] Halbach K.; Design of permanent magnet multipole magnets with oriented rare earth cobalt materials, Nuclear Instruments and Methods 169, 1980s 9/9

arxiv: v1 [physics.plasm-ph] 1 Feb 2008

arxiv: v1 [physics.plasm-ph] 1 Feb 2008 Field structure and electron life times in the MEFISTO Electron Cyclotron Resonance Ion Source arxiv:0802.0092v1 [physics.plasm-ph] 1 Feb 2008 Abstract M. Bodendorfer 1,2, K. Altwegg 2, H. Shea 1, P. Wurz